Security & Trust

Your data stays yours.

Transparent architecture, honest limitations, zero vendor lock-in.

What Klair is — and what it is not

Klair does not replace your tools. It adds a context layer on top of them. Your data stays in HubSpot, Apollo, Docusign, your email — exactly where it is today. Klair stores structured context alongside: who you contacted, when, what was discussed, what documents are linked, where deals stand. Your source tools remain the source of truth.

Without Klair

Your tools work in silos. Your AI starts from zero every session. You re-explain context manually, every time.

With Klair

Your AI has structured context at every session — contacts, actions, documents — without you repeating yourself.

Architecture

Klair does not connect to your other tools directly.

Your AI connects to Klair via MCP. You decide what gets stored there.

You→Interact with your AI (Claude, ChatGPT, Gemini…)
Your AI→Connects to Klair via MCP to read and write structured context
Klair→Stores context layer: contacts, actions, documents, communications
Your tools→HubSpot, Apollo, Docusign… remain your source of truth, unchanged

What you gain with Klair

Persistent context
Your AI knows who you worked on yesterday, what was said, where deals stand — without re-explaining anything.
Usecases — Free
Pre-built prompts for repetitive tasks

A library of pre-built prompts for simple, repetitive tasks: follow up with a contact, check pipeline status, draft a message, log an activity.

Instead of explaining the same task to your AI every time, you pick a usecase. It already knows the structure, the data sources, and the expected output. You can still customise it to your needs.

Skills — Paid
Full professional workflow bricks

Recruitment screening, contract review, sales sequences, onboarding automation.

A Skill orchestrates multiple tools, data sources and actions from your AI in a single, repeatable flow. Built by domain experts — or build and sell your own.

Communications centralised
Connect LinkedIn, WhatsApp, Telegram, email — every exchange attached to the right contact profile automatically.
Your own classification
Create your own tags and selectors to classify contacts, relationships and actions — exactly the way you work.

Your data, your rules

security.noDataLossLabel
Klair stores a context layer — your original data stays in your tools. Nothing is moved or deleted from your existing systems.
security.ownershipLabel
You own your data in Klair — always. Klair never claims rights over it.
security.exportLabel
Export all your Klair data at any time in standard formats (JSON, CSV).
security.deletionLabel
Delete any data stored in Klair at any time. Deletion is complete and permanent.
security.aiTrainingLabel
Your data is never used to train AI models — not Klair's, not anyone else's.
security.controlLabel
You decide what your AI is allowed to store in Klair — read-only or read-write, per data type.

Infrastructure & encryption

security.dataLocationLabel
Data stored in the European Union (EU).
security.encryptionRestLabel
All stored data encrypted at rest (AES-256).
security.encryptionTransitLabel
All data in transit encrypted (TLS 1.3).
security.tenantIsolationLabel
Your data is strictly isolated from other users. No cross-tenant access is technically possible.
security.auditLogsLabel
Every action your AI takes on your Klair data is logged — what was read, what was written, when.

GDPR & compliance

security.gdprLabel
Fully GDPR-compliant. Data stored in the EU. Right to access, correction and erasure at any time.
security.dpaLabel
GDPR-compliant Data Processing Agreement available on request.
security.subprocessorsLabel
Full list of subprocessors available on request.
security.rightForgottenLabel
Full deletion of account and all associated data within 30 days of request.

AI write access — what your AI can do

Your AI can read and write to your Klair context layer. Every write action is recorded in your audit log and reversible.

security.sensitiveActionsLabel
Actions with external consequences always require your explicit confirmation. Your AI cannot trigger them autonomously.
security.promptInjectionLabel
All data is validated before being written to Klair. A malicious email cannot instruct your AI to corrupt your context.

Vendor risk & lock-in

security.shutDownLabel

Your original data stays in your tools — HubSpot, Apollo, Docusign — untouched. Export your Klair context layer in open formats (JSON, CSV) at any time.

security.protocolDepLabel

Klair uses MCP — an open standard governed by the Linux Foundation, not by Klair. Supported by Anthropic, OpenAI, Google, Microsoft and Amazon.

security.noBgIntegrationsLabel

Klair maintains no background connections to your other tools. Your AI is the only client that accesses Klair.

Honest limitations

Klair is as complete as what you put in it

If some interactions happen outside your connected workflow, your AI's picture will be incomplete. Klair does not capture everything automatically.

AI summaries can be imprecise

When your AI summarises a conversation and writes it to Klair, nuance can be lost. Review AI-written entries for anything high-stakes.

Klair is a new product

We are early. We do not yet have SOC 2 Type II certification or third-party penetration test reports publicly available. We recommend starting with non-critical workflows and scaling as trust is established.